Differential fertility rates among cultural groups are, however, of significance only insofar as the difference is seen as having a significant political effect. In the United States little if any attention is paid to differential fertility rates among Baptists, Methodists, Presbyterians, and Episcopalians, but attention is given to the differentials between Protestant and Catholics and between Negroes and whites. But in most societies a differential in the growth rate of races receives attention since racial differences are generally considered more enduring and less assimilable than are differences in language or religion. In South Africa, for example, considerable attention is given to the higher fertility rates of Bantus, "coloureds," and Asians, compared to that of whites.* And though the Maoris constitute less than 10 percent of the population of New Zealand, there has been some concern in that country because the Maori birth rate is approximately twice that of the white population. "If the white people of New Zealand," wrote one extremist New Zealander, "do not awaken to their responsibilities and if the present trend of Maori population continues, this country will inevitably be populated mainly by a coloured race" (53). A similar concern with the higher fertility of a racial group can be found in the contemporary debate on race issues in Great Britain. In Britain, the growth in the size of the "coloured" (mainly Indian) population has become a strong argument for barring immigrants and for encouraging nonwhites to emigrate. Indeed, in all three cases, in South Africa, in New Zealand, and in Great Britain, the higher fertility rates of racial minorities have been used as a political argument for restrictive and selective immigration. Conflicts over the Census. When the representation of ethnic groups in the political system is explicitly based upon numbers, there is often a political struggle over control of the enumerating agency—the census department. Controversies often arise over such issues as which ethnic groups will be employed as enumerators, the reliability of the enumerators in finding and counting specific ethnic groups, whether certain social groups have been under-enumerated, and the criteria employed for indicating race, or religion, or language. In Lebanon, the census itself is such a delicate issue that the Lebanese government has been reluctant to take a census which records religion for fear that a change in the distribution of Muslims and Christians in the country would disturb the existing political balance within the government. In India, religion and language are enumerated but, with certain exceptions, caste is not, since the Indian government does not accept caste as a legitimate social institution. In fact, India can provide us with a number of examples of how the census can be employed by government or by social groups to change the numbers *According to Strauss (52), the natural population increase for whites in 1960 was 1.6 nercent. for Asians 2.8 oercent. for "coloureds" 3.3 oercent. Reliable data on thele argument, therefore, cuts both ways. less important.